WestminsterResearch http://www.westminster.ac.uk/westminsterresearch Investigating mitochondrial DNA relationships in Neolithic Western Europe through serial coalescent simulations Rivollat, M., Rottier, S., Couture, C., Pemonge, M.H., Mendisco, F., Thomas, M.G., Deguilloux, M.F. and Gerbault, P. This is an author's accepted manuscript of an article published in the European Journal of Human Genetics, 28 December 2016. The final definitive version is available online at: https://dx.doi.org/10.1038/ejhg.2016.180 The WestminsterResearch online digital archive at the University of Westminster aims to make the research output of the University available to a wider audience. Copyright and Moral Rights remain with the authors and/or copyright owners. Whilst further distribution of specific materials from within this archive is forbidden, you may freely distribute the URL of WestminsterResearch: ((http://westminsterresearch.wmin.ac.uk/). In case of abuse or copyright appearing without permission e-mail [email protected]
17
Embed
Western Europe through serial coalescent simulations ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
SHORT REPORT 1 2Title: 3Investigating mitochondrial DNA relationships in Neolithic Western Europe through 4serial coalescent simulations 5 6Running title: 7Genetic relationships in Neolithic western Europe 8 9 10Authors: 11Maïté Rivollat1, Stéphane Rottier1, Christine Couture1, Marie-Hélène Pemonge1, Fanny 12Mendisco1, Mark G. Thomas2, Marie-France Deguilloux1, Pascale Gerbault2,3,4 13 14Affiliations: 151 De la Préhistoire à l’Actuel, Culture, Environnement, Anthropologie – UMR 5199, 16CNRS, Université de Bordeaux, Bordeaux, France 172 UCL Research Department of Genetics, Evolution and Environment, Darwin building, 18Gower Street, London WC1E 6BT, UK 193 UCL Department of Anthropology, 14 Taviton Street, London WC1H 0BW, UK 204 Department of Life Sciences, University of Westminster, 115 New Cavendish Street, 21London W1W 6UW, UK 22 23Corresponding authors: 24Maïté Rivollat (email : [email protected] / [email protected]; Tel: +33 (0)5 2540 00 25 48; Fax: +33 (0)5 40 00 25 45); Pascale Gerbault (email: [email protected]; Tel: +44 (0) 207 679 4397; Fax: +44 (0) 207 679 7193) 27 28Conflict of interest: 29The authors have nothing to disclose, no conflict of interest. 30 31 32 33
1
Abstract: 34
Recent ancient DNA studies on European Neolithic human populations have provided 35
persuasive evidence of a major migration of farmers originating from the Aegean, 36
accompanied by sporadic hunter-gatherer admixture into early Neolithic populations, 37
but increasing towards the Late Neolithic. In this context, ancient mitochondrial DNA 38
(mtDNA) data collected from the Neolithic necropolis of Gurgy (Paris Basin, France), 39
the largest mtDNA sample obtained from a single archaeological site for the 40
Early/Middle Neolithic period, indicate little differentiation from farmers associated to 41
both the Danubian and Mediterranean Neolithic migration routes, as well as from 42
western European hunter-gatherers. To test whether this pattern of differentiation could 43
arise in a single unstructured population by genetic drift alone, we used serial coalescent 44
simulations. We explore female effective population size parameter combinations at the 45
time of the colonization of Europe 45 000 years ago and the most recent of the Neolithic 46
samples analyzed in this study 5 900 years ago, and identify conditions under which 47
population panmixia between hunter-gatherers/Early-Middle Neolithic farmers and 48
Gurgy cannot be rejected. In relation to other studies on the current debate of the origins 49
of Europeans, these results suggest increasing hunter-gatherer admixture into farmers’ 50
group migrating farther west in Europe. 51
52
Key words: genetic drift, European Neolithic, serial coalescent, ancient DNA, mtDNA 53
54
Introduction: 55
The introduction of farming into Europe around 8 600 years ago led to fundamental 56
changes in subsistence strategy and social organization, and left signatures of 57
2
population turnover1-4. It is widely believed that farming spread into Europe from the 58
Aegean along both Mediterranean and Danubian routes3,5. Recent archaeological6 and 59
palaeogenetic evidence1,2,4,5,7 indicate a crucial role for migration, with only sporadic 60
hunter-gatherer (HG) admixture into early Neolithic populations, but increasing towards 61
the Late Neolithic1,7,8. However, these local inferences still permit spatiotemporal 62
heterogeneity in HG admixture during the Neolithic in continental Europe. 63
In this context, the mtDNA diversity of the Gurgy "Les Noisats" site, located south of 64
the Paris Basin and dated from 7 000 to 6 000 years ago, is striking since descriptive 65
analyses9 indicated affinities not only with early farmers associated with both the 66
Danubian and Mediterranean migration routes but also with European HG. Notably, a 67
relatively lower differentiation between Gurgy and European HG (FST=0.08) was 68
observed when compared to other published levels of differentiation between Early 69
Neolithic farmers and HG (e.g. FST=0.0923 (ref. 10); FST=0.163 (ref. 2)). This suggests 70
complex admixture pattern between HG and farmer groups to shape Gurgy mtDNA 71
diversity. 72
Previous mtDNA studies2,10,11 have used serial coalescent simulations to test for genetic 73
continuity between HG, Neolithic farmers and extant DNA samples from the same 74
geographic region, and regularly concluded in genetic discontinuity between groups. 75
We used a similar approach to address if the observed level of mtDNA differentiation 76
between European HG, Neolithic farmer and Gurgy groups could be obtained under a 77
panmictic population model with various combinations of effective population sizes. 78
Our approach differs in three major aspects from previous studies2,10,11: first, we 79
grouped the ancient mtDNA sequences according to subsistence strategy (HG or 80
Neolithic farmers) and Neolithic context (Mediterranean/South-, Danubian/Central- or 81
3
Gurgy- farmers). Some of the sample groups are consequently contemporaneous and 82
can represent various regions. Second, we did not include modern population sample 83
into the comparison. Third, we extended the effective population size ranges used 84
previously2,11 towards the lower bound to explore further demographic scenarios. 85
86
MATERIAL AND METHODS 87
We compiled 282 available ancient mtDNA HVR-I sequences 88
(NC_012920.1:m.16024_16380; Table S1). Following Rivollat et al. 2015, ancient 89
mtDNA data were partitioned into 4 sample groups: (i) Gurgy Les Noisats necropolis 90
(hereafter referred to as “Gurgy”, n=39 sequences), (ii) Neolithic farmers from south 91
Europe (group “South-F”, n=56, partitioned into 4 chronological sub-groups), (iii) 92
Neolithic farmers from central Europe (“Central-F”, n=147, 5 sub-groups), and (iv) 93
hunter-gatherers (“HG”, n=40, 16 sub-groups). Chronological sub-groups were defined 94
according to both shared geographic location and median calibrated C14 dates (see 95
Figure 1 and Figure 2). As a test statistic that measures the level of population 96
differentiation, we calculated six pairwise FST between the four groups (Figure 3) with 97
ARLSUMSTAT version 3.5.1.2 (ref. 12). 98
Following previous studies2,11 we performed serial coalescent simulations under a single 99
panmictic population model with two demographic events: an initial colonization of 100
Europe 45 000 years ago of female effective population size NUP, followed by 101
exponential growth or decline to the Neolithic transition in Western Europe 5 900 years 102
cal. BP of female effective population size NN. Prior to NUP we assume an ancestral 103
female effective population size NA of 5 000, derived from the commonly used long-104
term effective human population size of 10 000 individuals outside Africa13 and 105
4
assuming a 1:1 female to male ratio. We explored 50 values for NUP ranging from 1 to 5 106
000 and 50 values for NN ranging from 10 to 100 000 (Table S2). We generated 50 000 107
mitochondrial genealogies of ancient HG and farmer sequences using fastsimcoal 108
version 2.5.1 (ref. 14) under each of the 2 500 NUP - NN combinations (Table S2). We 109
used a fixed mutation rate of 5×10-6/bp/generation (ref. 15), assuming a 25 years 110
generation time. These simulated genealogies were used to compute expected pairwise 111
FST values for the six sample comparisons (Figure 2). We recorded the proportion of 112
simulated FST values that were greater than those observed per FST and parameter 113
combination (Figure 3). 114
We also tested if the six observed pairwise FST values as well as eight within sample 115
group statistic values (number of segregating sites and of pairwise differences) could be 116
recovered from simulations under this simple model by performing an approximate 117
Bayesian computation (ABC) -related approach16 (see details in SI). We used the 118
rejection algorithm of the ‘abc’ package17 available in R to retain the parameter 119
combinations that generated simulated pairwise FST the closest to the 6 observed values. 120
Even though we provide some effective population size estimates, we caution against 121
over-interpretation since there is likely insufficient information in the data to make 122
precise estimates. 123
124
125
Results and Discussion 126
Analyses indicate that for the six pairwise population group comparisons, some NUP - 127
NN combinations can result in simulated differentiation greater than the one observed 128
(grey area on Figure 3). Notably, results show that we cannot reject the possibility that 129
5
European HG, South-F, Central-F and Gurgy were sampled from a single panmictic 130
population. Whereas these results may appear to contrast with previous studies that have 131
used serial coalescent simulations to address local mtDNA population continuity 132
between diachronic HG and farmers samples2,11, we highlight that our analyses do not 133
address ‘population continuity’ as defined in these studies. The grouping of diachronic 134
samples may artificially reduce the level of differentiation that would be observed in 135
case of significant mtDNA population structure. This grouping none-the-less allows us 136
to investigate the genetic relationships between set of lineage samples associated with 137
specific archaeological Neolithic contexts. 138
We confirmed that our panmictic population model generated simulated between and 139
within population group diversity values close to the observed using an ABC-rejection 140
algorithm (see SI and Figure S1). The 95% credible intervals estimated from the 141
retained simulations are [5 – 3500] NUP females and [200 – 7750] NN females. These 142
estimates concur with the observation that the parameter space for which a panmictic 143
population model may hold is rather narrow (Figure 3). Most NN values tested and 144
compatible with the level of mtDNA differentiation observed are relatively low (10 to 145
200 females for the South-F and Central-F comparison, Figure 3). Noteworthy, some 146
NUP - NN combinations imply a population decline that clearly contrasts with previous 147
studies based on modern DNA data which have inferred female effective population 148
size growth in Europe during the Holocene18. However, we were not constrained to 149
simulate population expansion, since we did not consider modern DNA data in our 150
analyses. Moreover, a Holocene population decline in Europe corroborates recent Y 151
chromosome data18 and various archaeological evidence support demographic 152
fluctuation of Neolithic populations19,20. 153
6
Our results indicate that a simple panmictic population model can account for the 154
mtDNA differentiation observed between European HG and Early/Middle Neolithic 155
farmers; a larger proportion of the HG - Gurgy explored parameter space failed to reject 156
panmixia. This result suggests increasing HG admixture into farmers’ group migrating 157
farther west in Europe. Similarly, we note that a larger proportion of the explored 158
parameter space fails to reject panmixia when comparing Gurgy and South-F than when 159
comparing Gurgy and Central-F. Thus, our results seem to support Gurgy as the most 160
ancient Neolithic sample studied so far with appreciable admixture between pre-161
Neolithic HG and Early/Middle Neolithic farmers from both streams of Neolithization 162
in Europe (with a suspected higher participation of Mediterranean farmers). 163
As with any model, the one we test here has a few assumptions that may not hold, e.g. 164
NA of female to male ratio of 1 (ref. 18) and no population structure in any of the four 165
groups5. Moreover, the panmictic population model proposed would need to be 166
compared against alternatives (e.g. ref. 11). Such a simple panmictic population model 167
nevertheless lays the ground for building more complex ones17. Notably, a serial 168
coalescent approach coupled with ABC would allow estimation of the possible 169
contribution of each of the three population groups (HG, Mediterranean and Central 170
Europe farmers) in shaping Gurgy mitochondrial diversity. 171
172
173
Acknowledgements: 174
This work was supported by a ministerial grant from the Research National Agency as a 175
program of prospects investments (ANR-10-LABX-52, DHP project; dir: SR; 176
Université Bordeaux 1, LaScArBx-ANR; 2012-14; and ANR-10-IDEX-03-02 to MR 177
7
for work in UCL) and a Leverhulme Programme grant to A.M. Migliano (UCL 178
Anthropology) and M.G. Thomas (RP2011-R-045). The authors acknowledge the use of 179
the UCL Legion High Performance Computing Facility (Legion@UCL), and associated 180
support services, in the completion of this work. 181
182
Supplementary information is available at European Journal of Human Genetics’ 183